U.S. EXPERIENCE WITH SPRINKLERS AND OTHER AUTOMATIC FIRE EXTINGUISHING EQUIPMENT

Transcription

1 U.S. EXPERIENCE WITH SPRINKLERS AND OTHER AUTOMATIC FIRE EXTINGUISHING EQUIPMENT John R. Hall, Jr. February 2010 National Fire Protection Association Fire Analysis and Research Division

2 U.S. EXPERIENCE WITH SPRINKLERS AND OTHER AUTOMATIC FIRE EXTINGUISHING EQUIPMENT John R. Hall, Jr. February 2010 National Fire Protection Association Fire Analysis and Research Division

3 Abstract Automatic sprinklers are highly effective elements of total system designs for fire protection in buildings. They save lives and property, producing large reductions in the number of deaths per thousand fires, in average direct property damage per fire, and especially in the likelihood of a fire with large loss of life or large property loss. When sprinklers are present in the fire area, they operate in 93% of all reported structure fires large enough to activate sprinklers, excluding buildings under construction. When they operate, they are effective 97% of the time, resulting in a combined performance of operating effectively in 91% of reported fires where sprinklers were present in the fire area and fire was large enough to activate sprinklers. In homes (including apartments), wet-pipe sprinklers operated effectively 96% of the time. When wet-pipe sprinklers are present in structures that are not under construction and excluding cases of failure or ineffectiveness because of a lack of sprinklers in the fire area, the fire death rate per 1,000 reported structure fires is lower by 83% for home fires, where most structure fire deaths occur, and the rate of property damage per reported structure fire is lower by 40-70% for most property uses. In homes (including apartments), wet-pipe sprinklers were associated with a 74% lower average loss per fire. Also, when sprinklers are present in structures that are not under construction and excluding cases of failure or ineffectiveness because of a lack of sprinklers in the fire area, 95% of reported structure fires have flame damage confined to the room of origin compared to 74% when no automatic extinguishing equipment is present. When sprinklers fail to operate, the reason most often given (53% of failures) is shutoff of the system before fire began. (All statistics are based on fires reported to U.S. fire departments, excluding buildings under construction.) Keywords: fire sprinklers; fire statistics; automatic extinguishing systems; automatic suppression systems Acknowledgements The National Fire Protection Association thanks all the fire departments and state fire authorities who participate in the National Fire Incident Reporting System (NFIRS) and the annual NFPA fire experience survey. These firefighters are the original sources of the detailed data that make this analysis possible. Their contributions allow us to estimate the size of the fire problem. We are also grateful to the U.S. Fire Administration for its work in developing, coordinating and maintaining NFIRS. For more information about the National Fire Protection Association, visit or call To learn more about the One-Stop Data Shop go to or call Copies of this report are available from: National Fire Protection Association One-Stop Data Shop 1 Batterymarch Park Quincy, MA osds@nfpa.org phone: NFPA No. USS14 Copyright 2010, National Fire Protection Association, Quincy, MA

4 Executive Summary Automatic sprinklers are highly effective and reliable elements of total system designs for fire protection in buildings. In , sprinklers operated in 93% of all reported structure fires large enough to activate sprinklers, excluding buildings under construction and buildings without sprinklers in the fire area. When sprinklers operate, they are effective 97% of the time, resulting in a combined performance of operating effectively in 91% of all reported fires where sprinklers were present in the fire area and fire was large enough to activate them. The combined performance for the more widely used wet pipe sprinklers is 92%, while for dry pipe sprinklers, the combined performance is only 79%. In homes (including apartments), wet-pipe sprinklers operated effectively 96% of the time. By comparison, combined performance is 60% for dry chemical systems, 79% for carbon dioxide systems, 81% for foam systems, and 88% for halogen systems. (Wet chemical systems may be included with dry chemical systems or with other special hazard systems.) These most current statistics are based on fires reported to U.S. fire departments, excluding buildings under construction and cases of failure or ineffectiveness because of a lack of sprinklers in the fire area and after some recoding between failure and ineffectiveness based on reasons given. When wet-pipe sprinklers are present in structures that are not under construction and excluding cases of failure or ineffectiveness because of a lack of sprinklers in the fire area, the fire death rate per 1,000 reported home structure fires is lower by 83% and the rate of property damage per reported structure fire is lower by 40-70% for most property uses. In homes (including apartments), wet-pipe sprinklers were associated with a 74% lower average loss per fire. Also, when sprinklers are present in structures that are not under construction and excluding cases of failure or ineffectiveness because of a lack of sprinklers in the fire area, 95% of reported structure fires have flame damage confined to the room of origin compared to 74% when no automatic extinguishing equipment is present. Of reported structure fires in health care properties, an estimated 57% showed sprinklers present, with higher percentages for hospitals (71%) and nursing homes (65%) and a much lower percentage for clinics and doctor s offices (28%). Sprinklers were also reported as present in half or more of all reported fires in laboratories (60%), manufacturing facilities (52%), theaters (50%), and prisons and jails (50%). In every other property use, more than half of all reported fires had no sprinklers. The few surveys that have been done of sprinkler presence in general, not limited to fires, have found that in any property group, the percentage of buildings with sprinklers is much higher than the percentage of reported fires with sprinklers present. Sprinklers apparently are still rare in many of the places where people are most exposed to fire, including educational properties, offices, most stores, and especially homes, where most fire deaths occur. There is considerable potential for expanded use of sprinklers to reduce the loss of life and property to fire. When sprinklers fail to operate, the reason most often given (53% of failures) was shutoff of the system before fire began, as may occur in the course of routine inspection maintenance. Other leading reasons were inappropriate system for the type of fire (20%), lack of maintenance (15%), and manual intervention that defeated the system (9%). Only 2% of sprinkler failures were attributed to component damage. U.S. Experience With Sprinklers, 2/10 i NFPA Fire Analysis and Research, Quincy, MA

5 When sprinklers operate but are ineffective, the reason usually had to do with an insufficiency of water applied to the fire, either because water did not reach the fire (43% of cases of ineffective performance) or because not enough water was released (31%). Other leading reasons were inappropriate system for the type of fire (12%), manual intervention that defeated the system (5%), and lack of maintenance (4%). Only 4% of cases of sprinkler ineffectiveness were attributed to component damage. When people are fatally injured in spite of the operation of wet-pipe sprinklers, the victims often had special vulnerabilities that are less often found with fatal victims of home fires in general. For example, 93% of fatal victims in home fires with wet-pipe sprinkler operation were located in the area of fire origin, where they could have suffered fatal injuries before sprinkler activation, compared to 53% of fatal home fire victims in general; 30% of fatal victims in home fires with wet-pipe sprinkler operation had their clothing on fire, compared to 7% of fatal home fire victims in general; 50% of fatal victims in home fires with wet-pipe sprinkler operation were age 65 or older, compared to 28% of fatal home fire victims in general; and 37% of fatal victims in home fires with wet-pipe sprinkler operation returned to the fire after escaping, compared to 19% of fatal home fire victims in general. U.S. Experience With Sprinklers, 2/10 ii NFPA Fire Analysis and Research, Quincy, MA

6 Table of Contents Executive Summary i Table of Contents iii List of Tables v Fact Sheet vii Before You Read the Report: Some Introductory Notes on Incident 1 Coding and Analysis Presence of Sprinklers and Other Automatic Extinguishing Equipment 3 Automatic Extinguishing Equipment Type 7 Automatic Extinguishing Equipment Reliability and Effectiveness 11 Automatic Extinguishing Equipment Impact 39 Other Issues 47 Concluding Points 53 Appendix A: How National Estimates Statistics Are Calculated 55 Appendix B: Sprinkler-Related Data Elements in NFIRS Appendix C: Multiple-Death Fires in Fully Sprinklered Properties 65 Appendix D: Selected Incidents 67 U.S. Experience With Sprinklers, 2/10 iii NFPA Fire Analysis and Research, Quincy, MA

7 U.S. Experience With Sprinklers, 2/10 iv NFPA Fire Analysis and Research, Quincy, MA

8 List of Tables Table 1. Presence of Sprinklers and Other Automatic Extinquishing Equipment 5 in Structure Fires Table 2. Table A. Table B. Type of Automatic Extinguishing Equipment Reported as Percentage of All Structure Fires Where Equipment Was Present and of Known Type, by Property Use 9 Non-Confined Fires With Areas of Origin That Could be Room Larger Than the Sprinkler Design Area for the Space 14 Combined Sprinkler Performance vs. Sprinkler Success in Confining Fire to Room of Origin 16 Table C. Reasons for Failure or Ineffectiveness as Percentages of All Cases of Failure 17 or Ineffectiveness, for All Structures and All Sprinklers Table D. Leading Areas of Origin for Fires in One- or Two- Family Dwellings 20 Table 3. Table 4. Automatic Extinguishing Equipment Reliability and Effectiveness, by Property Use 22 Reasons for Failure to Operate When Fire Was Large Enough to Activate Equipment and Equipment Was Present in Area of Fire 28 Table 5. Reasons for Ineffectiveness When Fire Was Large Enough to Activate 31 Equipment and Equipment Was Present in Area of Fire, by Property Use Table 6. Extent of Flame Damage, for Sprinklers Present vs. Automatic Extinguishing 34 Equipment Absent Table 7. Number of Sprinklers Operating 35 Table 8. Sprinkler Effectiveness Related to Number of Sprinklers Operating 37 Table 9. Estimated Reduction in Civilian Deaths per Thousand Fires Associated 43 With Wet Pipe Sprinklers, by Property Use Table 10. Characteristics of Fatal Victims When Wet Pipe Sprinklers Operate vs. 44 All Conditions Table 11. Estimated Reduction in Average Direct Property Damage per Fire 45 Associated With Wet Pipe Sprinklers Table E. Non-Fire Sprinkler Activations and Major Property Use Group 48 Table F. Non-Fire Sprinkler Activations by Likelihood of Water Release and 48 Major Property Use Group U.S. Experience With Sprinklers, 2/10 v NFPA Fire Analysis and Research, Quincy, MA

9 U.S. Experience With Sprinklers, 2/10 vi NFPA Fire Analysis and Research, Quincy, MA

10 One-Stop Data Shop Fire Analysis and Research Division 1 Batterymarch Park, Quincy, MA osds@nfpa.org U.S. Experience with Sprinklers Sprinklers save lives and protect property from fires. Compared to properties without automatic extinguishing equipment The death rate per fire in sprinklered homes is lower by 83%. For most property uses, damage per fire is lower by 40-70% in sprinklered properties. Flame damage was confined to the room of origin in 95% of fires in sprinklered properties vs. 74% in fires with no automatic extinguishing equipment. Damage per Fire With and Without Sprinklers, Eating or drinking Educational Health care* Home including Apartment Hotel or motel $12,000 $18,000 $7,000 $8,000 $3,000 $17,000 $4,000 $19,000 $9,000 $42,000 Without automatic extinguishing equipment With sprinklers Store or office $44,000 $26,000 $0 $20,000 $40,000 $60,000 $80,000 $100,000$120,000 *Health care refers to hospitals, nursing homes, clinics, doctor s offices, and mental retardation facilities. Sprinklers are reliable and effective. In reported structure fires large enough to activate them, sprinklers operated in 93% of fires in sprinklered properties. Wet pipe sprinklers operated in 95% of these fires vs. 83% for dry pipe sprinklers. In reported structure fires large enough to activate them, sprinklers operated and were effective in 91% of fires in sprinklered properties. Wet pipe sprinklers operated and were effective in 92% of fires vs. 79% for dry pipe sprinklers. NOTE: NFPA s Fire Sprinkler Initiative: Bringing Safety Home is a nationwide effort to encourage the use of home fire sprinklers and the adoption of fire sprinkler requirements for new construction. See Statistics are based on U.S. reported fires excluding buildings under construction. Sprinklered properties exclude properties with no sprinklers in fire area. U.S. Experience With Sprinklers, 2/10 vii NFPA Fire Analysis And Research, Quincy, MA

11 The graph below is based on the 7% of fires in sprinklered properties (roughly 1,000 fires per year) in which the sprinkler should have operated but did not. Reasons When Sprinklers Fail to Operate System shut off before fire 53% Inappropriate system for fire 20% Lack of maintenance 15% Manual intervention defeated system 9% Damaged component 2% 0% 10% 20% 30% 40% 50% 60% In fires where sprinklers operated, they were effective in 97% of the cases. The graph below is based on the other 3% (roughly 400 fires per year), in which the sprinkler was ineffective. Reasons When Sprinklers Are Ineffective Water did not reach fire 43% Not enough water released 31% Inappropriate system for fire 12% Manual intervention defeated system Damaged component Lack of maintenance 5% 4% 4% 0% 10% 20% 30% 40% 50% Usually only 1 or 2 sprinklers are required to control the fire. When wet pipe sprinklers operated, 89% of reported fires involved only 1 or 2 sprinklers. For dry pipe sprinklers, 74% involved only 1 or 2 sprinklers. Statistics are based on U.S. reported fires excluding buildings under construction. Sprinklered properties exclude properties with no sprinklers in fire area. U.S. Experience With Sprinklers, 2/10 viii NFPA Fire Analysis And Research, Quincy, MA

12 Before You Read the Report: Some Introductory Notes on Incident Coding and Analysis See Appendix A for general information on the statistical methodology and see Appendix B for a detailed overview of data elements related to automatic extinguishing equipment. Here are some important points on incident coding and analysis that apply to this report: Fires excluded from analysis Fires in buildings with reported structure status of under construction are excluded. No fire protection systems or features can be expected to perform as designed in a building that is under construction. Statistics on reliability, effectiveness, and performance exclude partial systems as identified by reason for failure and ineffectiveness equal to equipment not in area of fire. Not all partial systems will be so identified, and the codes and standards for this equipment do not require coverage in all areas. For example, concealed spaces and exterior locations may not be required to have coverage. Missing choices and misleading labels when coding presence or type of automatic extinguishing report The established generic name of automatic extinguishing equipment is misleading, because many if not most such equipment is designed to control fires and not to fully extinguish them. There is no code for wet chemical system, which was mandated as the type of non-waterbased system to be used in eating and drinking establishments shortly after the coding rules were set for NFIRS Version 5.0, the current version of the U.S. Administration s National Fire Incident Reporting System. 1 Wet chemical systems may be coded as dry chemical systems, foam systems, or other special hazard systems and are probably more common than all of these other systems. Fire extinguishers are not automatic equipment and should not be coded but sometimes are reported under any of several types of automatic extinguishing equipment. There was no way to code automatic extinquishing equipment as unknown during 1999 to 2003, although there was the option of leaving the field blank. During that period, the U.S. Fire Administration advised that unknowns should be reported as no equipment present. 2 This arrangement had the potential to severely understate the presence of automatic extinguishing equipment. However, the estimates for 2002 and 2003 are not substantially lower than either the pre-1999 estimates or the three years of estimates from 2004 and later. Therefore, this potential problem seems to have had little effect in practice. 1 NFIRS compiles fire incident and casualty reports from participating U.S. local fire departments. NFPA s national estimates are based on NFIRS data and estimated totals from the annual NFPA fire experience survey of U.S. fire departments. 2 U.S. Fire Administration, NFIRS Coding Questions, revised January 2, 2002, p.13. U.S. Experience With Sprinklers, 2/10 1 NFPA Fire Analysis and Research, Quincy, MA

13 Recoding of sprinkler performance based on reasons for failure or ineffectiveness. The coding of reasons for failure or ineffectiveness has been used in this analysis to recode system performance entries. Unknown reasons have been proportionally allocated to avoid the dubious alternative assumption that the coded performance is correct if no reason is given for the performance. If Performance = Not Effective And Reason = System shut off Not in area of fire If Performance = Failed to Operate And Reason = Not enough agent Agent didn t reach fire Not in area of fire Then Change to: Performance = Failed to operate Presence = No; Performance not applicable Then Change to: Performance = Not effective Performance = Not effective Presence = No; Performance not applicable Note that this recoding will not address partial sprinkler systems where there were sprinklers in part or all of the fire area unless the system is ineffective because of fire spread to or from uncovered areas. U.S. Experience With Sprinklers, 2/10 2 NFPA Fire Analysis and Research, Quincy, MA

14 Presence of Sprinklers and Other Automatic Extinguishing Equipment Of reported structure fires in health care properties, an estimated 57% showed sprinklers present, with higher percentages for hospitals (71%) and nursing homes (65%) and a much lower percentage for clinics and doctor s offices (28%). Sprinklers were also reported as present in half or more of all reported fires in laboratories (60%), manufacturing facilities (52%), theaters (50%), and prisons and jails (50%). In every other property use, more than half of all reported fires had no sprinklers. In , only 7% of reported structure fires had any type of automatic extinguishing equipment present. By , this percentage had risen by about half, to 10%. Before 1999, the type of automatic extinguishing equipment was not reported, and so it is not possible to show the trend in sprinkler presence. It is possible to show the trend in presence of automatic extinguishing equipment generally and to show how sprinkler presence compares to automatic extinguishing equipment presence in the most recent years. See Table 1 for percentage of reported structure fires, excluding buildings under construction, in which automatic extinguishing equipment was present for the year groups of and Table 1 also shows percentage of fires with any type of sprinkler reported present for 2003 to The following properties where large numbers of people routinely are present show less than one-third of reported fires in properties with sprinklers present in : Every type of public assembly property except theaters Educational properties Clinics and doctor s offices Homes including apartments Every type of store or office except department stores Most fires in storage properties are not in warehouses but are in garages, barns, silos, and small outbuildings. It is these types of buildings that drive the very low percentage of reported fires with automatic extinguishing equipment in all storage properties combined. In , sprinklers were reported in only 5% of fires in homes (including apartments). Clearly, there is great potential for expanded installation. The 2007 American Housing Survey included a question about sprinkler presence in homes. 4 The survey indicated 3.9% of occupied year-round housing units had sprinklers. A much smaller percentage of single family homes had sprinklers as compared to multi-unit housing. Sprinklers were present in: 3 Some fires after 1999 are coded as confined fires, which are fires confined to cooking vessel, chimney or flue, furnace or boiler, incinerator, commercial compactor, or trash receptacle. Confined fires permit limited reporting with most data fields not required and usually left blank. Confined fires permit limited reporting with most data fields not required and usually left blank. Confined fires combine with very low sprinkler usage to make estimates for one- and two-family dwellings too volatile and uncertain to list separately, and so estimates are provided only for all homes combined 4 American Housing Survey 2007, U.S. Department of Commerce and U.S. Department of Housing and Urban Development, September 2008, Table 1C-4, 2-4, and U.S. Experience With Sprinklers, 2/10 3 NFPA Fire Analysis and Research, Quincy, MA

15 1.5% of single family detached homes, 1.9% of single family homes, whether detached or attached, 10.6% of all housing units in multi-unit buildings, 2.9% of housing units in buildings with 2-4 units, 5.8% of housing units in buildings with 5-9 units, 12.1% of housing units in buildings with units, 16.3% of housing units in buildings with units, and 27.3% of housing units in buildings with 50 or more units. Sprinklers are installed in 13.0% of housing units in buildings that were constructed no more than four years ago. This is more than triple the percentage for all housing units. No statistics are provided on sprinkler installation specifically in recently constructed single family homes, but detached single-family homes are a larger share of recently built housing units than of total housing units (70% vs. 63%). This strongly suggests that single family homes are part of the recent jump in sprinkler installation. Sprinkler presence percentages are higher in the West region than in other regions and lower in rural areas than in non-rural areas. To underscore the principal finding, more than 1 million single family detached dwellings now have fire sprinklers. The Home Fire Sprinkler Coalition, formed in 1996, developed a variety of educational materials about the benefits of home fire sprinklers. These materials address common questions and misconceptions. They may be accessed through their web site Because sprinkler systems are so demonstrably effective, they can make a major contribution to fire protection in any property. NFPA 101, Life Safety Code; NFPA 1, Fire Code; and NFPA 5000, Building Construction and Safety Code, have required sprinklers in all new one- and two-family dwellings, all nursing homes, and many nightclubs since the 2006 editions. The 2009 edition of the International Residential Code, also added requirements for sprinklers in one- and two-family dwellings, effective January This protection can be expected to increase in areas that adopt and follow these codes. NFPA is supporting adoption of these requirements through its Fire Sprinkler Initiative (see The few surveys that have been done of sprinkler presence in general, not limited to fires, have found that in any property group, the percentage of buildings with sprinklers is much higher than the percentage of reported fires with sprinklers present. Sprinklers apparently are still rare in many of the places where people are most exposed to fire, including educational properties, offices, most stores, and especially homes, where most fire deaths occur. There is considerable potential for expanded use of sprinklers to reduce the loss of life and property to fire. As with detection/alarm systems and all other fire protection features, in property classes where sprinklers are not required, they will tend to go first into the properties that can afford them most, not the high-risk fire-prone properties that would benefit most from their presence. U.S. Experience With Sprinklers, 2/10 4 NFPA Fire Analysis and Research, Quincy, MA

16 Table 1. Presence of Sprinklers and Other Automatic Extinguishing Equipment in Structure Fires, vs Number of Structure Fires With Equipment Present and Percentage of Total Structure Fires in Property Use Any Automatic Extinguishing Equipment Any Sprinkler Property Use All public assembly 4,380 (26%) 7,650 (49%) 3,040 (19%) Fixed-use amusement place 150 (18%) 170 (29%) 150 (24%) Variable-use amusement place 140 (16%) 270 (22%) 260 (22%) Religious property 90 (5%) 280 (15%) 270 (14%) Library or museum 110 (28%) 190 (28%) 180 (28%) Eating or drinking establishment 3,240 (29%) 4,730 (58%) 1,380 (17%) Passenger terminal 60 (35%) 180 (28%) 110 (16%) Theater 110 (35%) 140 (51%) 140 (50%) Educational property 1,820 (24%) 2,250 (34%) 2,010 (31%) Health care property 4,400 (68%) 4,010 (61%) 3,770 (57%) Nursing home 2,060 (76%) 2,060 (70%) 1,910 (65%) Hospital 1,650 (74%) 1,210 (77%) 1,110 (71%) Clinic or doctor s office 70 (29%) 200 (28%) 200 (28%) Prison or jail 430 (19%) 290 (51%) 290 (50%) All residential 11,110 (3%) 26,980 (8%) 25,820 (7%) Home (including apartment) 8,440 (2%) 21,110 (5%) 20,130 (5%) Hotel or motel 1,690 (35%) 1,900 (48%) 1,790 (45%) Dormitory or barracks 620 (29%) 1,670 (46%) 1,550 (42%) Rooming or boarding home 230 (17%) 970 (33%) 950 (32%) Board and care home NA (NA) 900 (43%) 790 (38%) Store or office 5,230 (21%) 6,090 (30%) 4,660 (23%) Grocery or convenience store 1,190 (27%) 2,030 (44%) 1,010 (22%) Laundry or dry cleaning or 310 (13%) 350 (19%) 340 (18%) other professional service Service station or motor 230 (6%) 230 (10%) 170 (7%) vehicle sales or service Department store 1,100 (52%) 610 (43%) 560 (39%) Office 1,470 (25%) 1,210 (32%) 1,170 (31%) Laboratory 120 (48%) 110 (65%) 100 (60%) Manufacturing facility 6,400 (50%) 4,070 (57%) 3,740 (52%) All storage 1,090 (3%) 950 (4%) 920 (4%) Warehouse excluding cold 740 (22%) 510 (38%) 510 (38%) storage** All structures 37,100 (7%) 53,940 (10%) 44,310 (9%) NA Category not defined in fire incident data prior to *Also includes development disability facilities. In , this category also includes care of physically inconvenienced and excludes doctor s office and care of aged facilities without nursing staff. **In , includes general warehouse, textile storage, processed food storage except cold storage and storage of wood, paper, plastics chemicals, and metals. Notes: These are structure fires reported to U.S. municipal fire departments and so exclude fire reported only to Federal or state agencies or industrial fire brigades. Post-1998 estimates are based only on fires reported in Version 5.0 of NFIRS and include fires reported as confined fires. Estimates are not shown for because of lower participation in NFIRS Version 5.0 in those years. After 1998, buildings under construction are excluded. Source: NFIRS and NFPA survey. U.S. Experience With Sprinklers, 2/10 5 NFPA Fire Analysis and Research, Quincy, MA

17 U.S. Experience With Sprinklers, 2/10 6 NFPA Fire Analysis and Research, Quincy, MA

18 Automatic Extinguishing Equipment Type In reported fires, most automatic extinguishing equipment is recorded as sprinklers, and most sprinklers are wet pipe sprinklers. Table 2 shows the percentage of non-confined and confined fires, excluding buildings under construction, by type of automatic extinguishing equipment for each of the major property groups and some subgroups. 5 Percentage calculations are based only on fires where automatic extinguishing equipment presence and type were known and reported. In Version 5.0 of NFIRS, if multiple systems are present, the system coded is supposed to be the one system designed to protect the hazard where the fire started. Some type of sprinklers were present in 82% of fires where automatic extinguishing equipment was present. Wet pipe sprinklers accounted for 73% of all systems and so outnumbered dry pipe sprinklers by roughly 10-to-1. The major property class with the largest share for dry pipe sprinklers was storage, where dry pipe sprinklers accounted for 20% of the systems cited. Cold storage was the only property class for which dry pipe sprinklers constituted a majority (in this case, 53%) of systems cited. For public assembly properties, there was a 40% to 60% split between sprinklers and other types of automatic extinguishing equipment, respectively. Dry chemical systems accounted for 40% of the systems present. Eating or drinking establishments (the dominant part of public assembly) had a 29% to 71% split between sprinkler systems and other types of automatic extinguishing equipment, respectively. Dry chemical systems accounted for 47% of total systems in eating or drinking establishments, compared to a 29% share for all sprinklers combined. Note that wet chemical systems have no clearly identified equipment type category but have been the mandated type of system for eating and drinking establishments for roughly a decade. It seems likely that most of the dry chemical systems reported are either wet chemical systems or dry chemical extinguishers, which should not be reported as any type of automatic equipment. Public assembly properties, especially eating and drinking establishments, have the highest percentages for both dry chemical systems (40% and 47%, respectively) and other special hazard systems (11% and 12%, respectively), both of which probably are dominated by wet chemical systems, for which there is no labeled category. Roughly ten years ago, the applicable standards for eating and drinking establishments required that dry chemical systems be replaced by wet chemical systems. It seems likely that some wet chemical systems will be coded as other special hazard systems and some will be coded as dry chemical systems, the latter being the well-defined equipment type closest to a wet chemical system. It would be useful to have a better sense of what kind of equipment is coded as other special hazard systems. There are some types of automatic extinguishing equipment that do not fit exactly into any of the defined categories, such as equipment using wet chemicals. It is also 5 Some fires after 1999 are coded as confined fires, which are fires confined to cooking vessel, chimney or flue, furnace or boiler, incinerator, commercial compactor, or trash receptacle. Confined fires permit limited reporting with most data fields not required and usually left blank. Confined fires combine with very low sprinkler usage to make estimates for one- and two-family dwellings too volatile and uncertain to list separately, and so estimates are provided only for all homes combined U.S. Experience With Sprinklers, 2/10 7 NFPA Fire Analysis and Research, Quincy, MA

19 possible that some fires will be coded as other special hazard system when they really involved automatic extinguishing equipment of one of the defined types. The category also could be used for some devices that are not automatic and so should not be coded as automatic extinguishing equipment present, such as portable extinguishers. Some insight into what is being coded under other special hazard systems comes from a check of uncoded narratives for the three restaurant fires in recent years in Minnesota where such equipment was reported. (The narratives on these fires were part of a data set provided for a special analysis described on p. 49.) One fire involved a wet chemical system, and another involved an undefined hood system, which could have involved wet or dry chemical agents. The third fire involved use of portable extinguisher and should not have been coded as automatic extinguishing equipment present. U.S. Experience With Sprinklers, 2/10 8 NFPA Fire Analysis and Research, Quincy, MA

20 Table 2. Type of Automatic Extinguishing Equipment Reported as Percentage of All Fires Where Equipment Was Present and of Known Type, by Property Use Structure Fires Reported to U.S. Fire Departments Property Use Fires per year with any automatic extinguishing equipment All sprinklers Wet pipe sprinklers Dry pipe sprinklers Other sprinklers* All public assembly 7,650 40% 34% 2% 4% Fixed-use amusement place % 82% 3% 1% Variable-use amusement place % 84% 13% 0% Religious property % 86% 4% 7% Library or museum % 92% 5% 2% Eating or drinking establishment 4,730 29% 23% 2% 4% Passenger terminal % 35% 22% 1% Theater % 95% 3% 1% Educational property 2,250 89% 80% 6% 3% Health care property** 4,010 94% 81% 12% 1% Nursing home 2,060 93% 78% 15% 1% Hospital 1,210 92% 85% 6% 1% Clinic or doctor s office % 95% 3% 1% Prison or jail % 87% 10% 2% All residential 26,980 96% 87% 7% 2% Home (including apartment) 21,110 95% 86% 6% 3% Hotel or motel 1,900 94% 85% 6% 3% Dormitory or barracks 1,670 93% 77% 14% 1% Rooming or boarding house % 88% 11% 0% Board and care home % 82% 7% 0% Store or office 6,090 77% 67% 7% 3% Grocery or convenience store 2,030 50% 44% 3% 3% Laundry or dry cleaning or other % 85% 9% 1% professional service Service station or motor vehicle % 70% 5% 1% sales or service Department store % 78% 12% 1% Office 1,210 97% 85% 7% 4% Laboratory % 69% 2% 21% Manufacturing facility 4,070 92% 79% 10% 3% All storage % 75% 20% 2% Warehouse excluding cold storage % 82% 15% 1% All structures*** 53,940 82% 73% 7% 3% * Includes deluge and pre-action sprinkler systems and may include sprinklers of unknown or unreported type. ** Nursing home, hospital, clinic, doctor s office, or development disability facility *** Includes some property uses that are not shown separately. Note: These are based on structure fires reported to U.S. municipal fire departments in NFIRS Version 5.0 and so exclude fires reported only to Federal or state agencies or industrial fire brigades. Row totals are shown in the leftmost column of percentages, and sums may not equal totals because of rounding error. In Version 5.0 of NFIRS, if multiple systems are present, the system coded is supposed to be the one system designed to protect the hazard where the fire started. This field is not required if the fire did not begin within the designed range of the system. Buildings under construction are excluded. Source: NFIRS and NFPA survey. U.S. Experience With Sprinklers, 2/10 9 NFPA Fire Analysis and Research, Quincy, MA

21 Table 2. (Continued) Type of Automatic Extinguishing Equipment Reported as Percentage of All Fires Where Equipment Was Present and of Known Type, by Property Use Structure Fires Reported to U.S. Fire Departments Property Use All systems other than sprinklers Dry chemical system* Carbon dioxide (CO2) system Halogen type system* Foam system Other special hazard system* All public assembly 60% 40% 3% 3% 4% 11% Fixed-use amusement place 15% 14% 0% 0% 0% 1% Variable-use amusement place 3% 2% 0% 0% 0% 1% Religious property 3% 2% 0% 0% 0% 1% Library or museum 1% 1% 0% 0% 0% 0% Eating or drinking establishment 71% 47% 3% 3% 5% 12% Passenger terminal 42% 41% 0% 0% 0% 1% Theater 1% 1% 0% 0% 0% 0% Educational property 11% 8% 0% 0% 1% 2% Health care property** 6% 4% 1% 0% 0% 1% Nursing home 7% 5% 1% 0% 0% 0% Hospital 8% 5% 0% 0% 0% 3% Clinic or doctor s office 2% 0% 0% 0% 0% 1% Prison or jail 2% 1% 0% 0% 0% 0% All residential 4% 2% 0% 0% 0% 2% Home (including apartment) 5% 2% 0% 0% 0% 2% Hotel or motel 6% 2% 0% 0% 0% 4% Dormitory or barracks 7% 5% 0% 0% 1% 1% Rooming or boarding home 2% 1% 0% 0% 0% 0% Board and care home 11% 4% 0% 0% 4% 3% Store or office 23% 15% 2% 1% 2% 4% Grocery or convenience store 50% 31% 4% 1% 6% 7% Laundry or dry cleaning 5% 0% 0% 0% 0% 4% Service station or motor 24% 20% 0% 1% 0% 2% vehicle sales or service Department store 9% 8% 0% 0% 0% 0% Office 3% 1% 1% 0% 0% 0% Laboratory 8% 2% 4% 1% 0% 1% Manufacturing facility 8% 2% 4% 0% 0% 1% All storage 3% 1% 0% 0% 0% 2% Warehouse excluding cold storage 1% 0% 0% 0% 0% 0% Cold storage 0% 0% 0% 0% 0% 0% All structures*** 18% 11% 1% 1% 1% 4% * Dry chemical system may include wet chemical systems, because there is no category designated for wet chemical systems. Halogen type system includes non-halogenated suppression systems that operate on the same principle. Other special hazard system may include automatic extinguishing systems that are known not to be sprinklers but otherwise are of unknown or unreported type. ** Nursing home, hospital, clinic, doctor s office, or development disability facility. *** Includes some property uses that are not shown separately. Note: These are based on structure fires reported to U.S. municipal fire departments in NFIRS Version 5.0 and so exclude fires reported only to Federal or state agencies or industrial fire brigades. Row totals are shown in the leftmost column of percentages, and sums may not equal totals because of rounding error. In Version 5.0 of NFIRS, if multiple systems are present, the system coded is supposed to be the one system designed to protect the hazard where the fire started. This field is not required if the fire did not begin within the designed range of the system. Building under construction are excluded. Source: NFIRS and NFPA survey. U.S. Experience With Sprinklers, 2/10 10 NFPA Fire Analysis and Research, Quincy, MA

22 Automatic Extinguishing Equipment Reliability and Effectiveness In order to estimate the reliability and effectiveness of any type of automatic extinguishing equipment, the database must first be edited to remove fires, buildings, and systems where operation cannot be expected, such as small fires, buildings under construction, and partial installations. Table 3 shows the percentage of non-confined and confined structure fires, excluding buildings under construction and incidents with partial systems not in area of fire, where fires were too small to activate operational automatic extinguishing equipment. Table 3 also shows, for fires large enough to activate equipment, the percentage of fires where equipment operated, the percentage of operating equipment cases where equipment was effective, and the percentage of fires where equipment operated effectively. This is shown for: All sprinklers Wet pipe sprinklers Dry pipe sprinklers Dry chemical systems (which probably includes and may be dominated by wet chemical systems and may include some miscoded portable extinguishers), Carbon dioxide systems (which may include some wet chemical systems and some miscoded portable extinguishers), Foam systems (which may include some wet chemical systems and some miscoded portable extinguishers), and Halogen systems (which may include some wet chemical systems and some miscoded portable extinguishers). Property use classes are shown only if they accounted for at least 100 projected fires per year with the specific type of automatic extinguishing equipment present. For most property use groups and most types of automatic extinguishing equipment, the majority of reported fires were too small to activate operational equipment. When automatic extinguishing equipment was present, the percentages of fires too small to activate operating equipment, based on overall reported structure fires, were as follows: 65% for all sprinklers, 65% for wet pipe sprinklers, 70% for dry pipe sprinklers, 61% for dry (or possibly wet) chemical systems, 43% for carbon dioxide systems, 66% for foam systems, and 59% for halogen systems. Sprinklers in the area of fire failed to operate in only 7% of reported structure fires large enough to activate sprinklers. Failure rates are equal to 100% minus the percentage of systems that operated, which is the percentage shown in Table 3A. The other estimated failure rates corresponding to percentage operating rates shown in Table 3A are: 5% for wet pipe sprinklers, U.S. Experience With Sprinklers, 2/10 11 NFPA Fire Analysis and Research, Quincy, MA

23 17% for dry pipe sprinklers, 26% for dry (or possibly wet) chemical systems, 17% for carbon dioxide systems, 3% for foam systems, and 4% for halogen systems. For major property classes and sprinklers, the estimated failure rates range from a low of 4% for residential properties, public assembly properties and stores and offices to a high of 32% for educational properties and 29% for storage properties. The estimated failure rates for wet pipe sprinklers specifically were 25% for educational properties and 16% for storage properties. For sprinklers that operated, their performance was deemed effective in 97% of the cases. For all confined or non-confined fires large enough to activate sprinklers, excluding buildings under construction, sprinklers operated effectively 91% of the time. The percentages of effective operation for all structures were as follows for other types of automatic extinguishing equipment: 92% for wet pipe sprinklers, 79% for dry pipe sprinklers, 60% for dry (or possibly wet) chemical systems, 79% for carbon dioxide systems, 81% for foam systems, and 88% for halogen systems. Wet pipe sprinklers are both much more reliable than dry pipe sprinklers (95% vs. 83%) and slightly more effective when they operate (98% vs. 95%), resulting in a much higher percentage of effective operation (92% vs. 79%). Operating effectiveness is much lower for dry (or possibly wet) chemical systems than for any other type of automatic extinguishing equipment (60% vs %) and is especially low (51%) for eating or drinking establishments, which account for most of the fires reported with this type of equipment. Eating or drinking establishments also account for most fires reported with carbon dioxide, foam, or halogen systems. These installations may all include a high proportion of misclassified wet chemical systems or portable extinguishers, because carbon dioxide, foam, and halogen systems are rarely appropriate for eating or drinking establishments. A disadvantage of measuring automatic extinguishing equipment effectiveness by judgments made in incident reports is the ambiguity and subjectivity of the criterion of effective, which has never been precisely defined, let alone supported by an operational assessment protocol that could be executed consistently by different people. Also, confined fires usually have these details unreported, and so their few fires with details reported will be weighted far more heavily, after allocation of unknowns, than will non-confined fires. The majority of sprinkler failures occurred because the system was shut off. Table 4 provides the percentages of reasons for failure, after recoding, by type of automatic extinguishing system and property use. Other or unclassified reason for failure is treated as an unknown and allocated. U.S. Experience With Sprinklers, 2/10 12 NFPA Fire Analysis and Research, Quincy, MA

24 For all types of sprinklers combined: 53% of failures to operate were attributed to the equipment being shut off, 20% were because the equipment was inappropriate for the type of fire, 15% were because of lack of maintenance, 9% were because manual intervention defeated the equipment, and 2% were because a component was damaged. If manual intervention occurs before fire begins, one would expect that to be coded as system shut off before fire. If manual intervention occurs after sprinklers operate, one would expect that to constitute ineffective performance, not failure to operate. What is left is manual intervention after fire begins but before sprinklers operate, but we do not know whether that is the only condition associated with this coding. Only 2% were because of a failing of the equipment rather than a failing of the people who designed, selected, maintained, and operated the equipment. If these human failings could be eliminated, the overall sprinkler failure rate would drop from the estimated 7% of reported fires to less than 0.2%. That is the kind of sprinkler failure rate reported by Marryatt 6 for Australia and New Zealand, where high standards of maintenance are reportedly commonplace. Training can sharply reduce the likelihood of three other causes of failure system defeating due to manual intervention, lack of maintenance, and installation of the wrong system for the hazard. Most cases of sprinkler ineffectiveness were because water did not reach the fire (43%) or because not enough water was released (31%). Table 5 provides distributions of reasons for ineffectiveness, by property class and type of automatic extinguishing equipment. In Table 5, two of the reasons for ineffectiveness are (extinguishing) agent did not reach the fire and not enough (extinguishing) agent was released. For sprinklers, the agent is water. In addition to the two reasons cited, other reasons for sprinkler ineffectiveness for all structures were inappropriate equipment for the type of fire (12%), defeating due to manual intervention (5%), damage to a system component (4%), and lack of maintenance (4%). There are a number of different ways in which water may not reach the fire. One is shielded fires such as rack storage in a property with ceiling sprinklers only. Another is fire spread above exposed sprinklers, through unsprinklered concealed spaces, or via exterior surfaces. Another reason would be a deep-seated fire in bulk storage. A different kind of problem would be droplet sizes that are too small to penetrate the buoyant fire plume and reach the seat of the fire. Insufficient water can be released if there are problems with the system s water supply. This reason for ineffectiveness can also overlap with other reasons, such as inappropriate equipment (if, for example, the hazard has changed under the equipment and now requires a higher water flow density than is provided by the now inappropriate equipment) and defeating by manual intervention (if, for example, the sprinklers are turned off prematurely so that insufficient water reaches the fire). Insufficient water also could be one of the reasons that could be cited if a flash 6 H.W. Marryatt, Fire: A Century of Automatic Sprinkler Protection in Australia and New Zealand, , 2 nd edition, Victoria, Australia: Australian Fire Protection Association, U.S. Experience With Sprinklers, 2/10 13 NFPA Fire Analysis and Research, Quincy, MA

25 fire or a fire with several points of origin overwhelms the system or if an explosion reduces the water flow but does not cause complete system failure. Reasons for ineffectiveness are different for wet pipe sprinklers and dry pipe sprinklers, with dry pipe sprinklers having 60% of cases attributed to not enough water released compared to 25% for wet pipe sprinklers. Because the design of dry pipe sprinklers assures a delayed release of water, it is not surprising that when such systems are ineffective, an insufficiency of water is usually involved. Even a well-maintained, complete, appropriate system requires the support of a well-considered integrated design for all the other elements of the building s fire protection. Unsatisfactory sprinkler performance can result from an inadequate water supply or faulty building construction. More broadly, unsatisfactory fire protection performance can occur if the building s design does not address all five elements of an integrated system slowing the growth of fire, automatic detection, automatic suppression, confining the fire, and occupant evacuation. Effectiveness should be measured relative to the design objectives for a particular system. For most rooms in most properties, sprinklers are designed to confine fire to the room of origin. Table A. Non-Confined Fires With Areas of Origin That Could Be Room Larger Than the Sprinkler Design Area for the Space, as Percent of Total Non-Confined and Confined Structure Fires for Buildings Not Under Construction and With Sprinklers in Fire Area Percentage of Structure Fires Reported to U.S. Fire Departments Property Use Large Assembly Area (At Least 100 People) Sales, Showroom or Performance Area Storage Room, Area, Tank or Bin Shipping, Receiving or Loading Area Unclassified Storage Area All Areas Combined Eating or drinking 1.0% 0.2% 1.7% 0.2% 1.2% 4.3% establishment Public assembly excluding eating 2.4% 0.7% 0.9% 0.2% 0.8% 5.0% or drinking establishment Educational 1.3% 0.3% 0.6% 0.0% 0.7% 2.9% Health care property* 0.1% 0.0% 0.5% 0.0% 0.3% 0.9% Home (including apartment) 0.0% 0.0% 0.2% 0.0% 0.2% 0.4% Hotel or motel 1.6% 0.0% 0.3% 0.0% 0.5% 1.4% Store or office 0.1% 5.7% 2.2% 2.1% 2.2% 12.2% Manufacturing facility 0.1% 0.0% 2.5% 2.1% 1.7% 6.4% Warehouse excluding cold 0.1% 0.3% 5.0% 12.4% 9.9% 27.7% storage * Hospital, clinic, doctor s office, nursing home and development disability facility. Note: Percentages are defined as non-confined fires with indicated area of origin divided by total non-confined and confined fires with any area of origin. Percentages sum left to right and may not equal totals in last column because of rounding. Fires reported as confined fires are excluded from the numerator because such fires could not be large enough to exceed the sprinkler design area. Statistics are based on structure fires reported to U.S. municipal fire departments and so exclude fires reported only to Federal or state agencies or industrial fire brigades. Statistics exclude buildings under construction and fires with sprinklers not in fire area reported as reason for failure or ineffectiveness of automatic extinguishing equipment. Source: NFIRS and NFPA survey. U.S. Experience With Sprinklers, 2/10 14 NFPA Fire Analysis and Research, Quincy, MA